Diffusion-based speech enhancement with a weighted generative-supervised learning loss
This is an incremental improvement for speech enhancement systems, potentially improving audio quality in applications like hearing aids or communication devices.
The paper tackles the problem of inefficient incorporation of noisy speech conditioning in diffusion-based speech enhancement by augmenting the diffusion training objective with a supervised mean squared error loss. Experimental results show this approach is effective, though no concrete numbers are provided.
Diffusion-based generative models have recently gained attention in speech enhancement (SE), providing an alternative to conventional supervised methods. These models transform clean speech training samples into Gaussian noise centered at noisy speech, and subsequently learn a parameterized model to reverse this process, conditionally on noisy speech. Unlike supervised methods, generative-based SE approaches usually rely solely on an unsupervised loss, which may result in less efficient incorporation of conditioned noisy speech. To address this issue, we propose augmenting the original diffusion training objective with a mean squared error (MSE) loss, measuring the discrepancy between estimated enhanced speech and ground-truth clean speech at each reverse process iteration. Experimental results demonstrate the effectiveness of our proposed methodology.